The development of an efficacious preventive HIV vaccine is a global priority in public health. This application proposes to conduct research in biostatistical methods for preventive HIV vaccine efficacy trials, which must be appropriately designed and analyzed to develop effective HIV vaccines as quickly as possible. The proposed methods evaluate the effects of vaccination on the risk of infection with various HIV strains and on biomarker outcomes measured after infection. The first aim is to develop methods for assessing the impact of HIV variation on vaccine efficacy to prevent HIV infection, including: (1) survival analysis methods to evaluate the relationship between vaccine efficacy and the genotypic or phenotypic distance of an exposing HIV strain to the prototype -IIV strain(s) represented in the tested vaccine, and (2) high-dimensional data genome scanning methods to identify short HIV peptide regions (i.e., 8-12 contiguous amino acid positions) at which the peptide sequences from infected vaccine recipients tend to be more divergent from the prototype sequence than those from infected placebo recipients. An HIV vaccine may modify viral load or other biomarkers in vaccinees who become infected, which may imply important vaccine benefits to reduce transmission and disease progression. The second aim is to develop methods for assessing the effect of vaccination on biomarker outcomes measured after infection, that appropriately account for selection bias that may arise because the analyzed groups are selected after randomization, and that appropriately account for potent treatment of some infected participants. To complement the second objective, the third aim is to develop models of longitudinal biomarker processes in infected participants (e.g., viral load or CD4 cell count profiles) that flexibly accommodate time-varying effects of covariates (e.g., immune responses). The fourth aim is to develop accurate simultaneous confidence interval procedures for assessing time-varying effects of vaccination to prevent infection or post-infection outcomes. [unreadable] [unreadable]